دانلود کتاب Cybernetics 2.0 – نظریه عمومی سازگاری و هموستاز در مغز و بدن

The human brain has about 100 billion neurons. Each neuron has about 10 thousand synapses. The total number of synapses is in the trillions. A synapse is the link that connects one neuron to another. The strength of the synaptic coupling can vary and, according to Hebb, synaptic change is associated with learning. What is the mechanism or algorithm that nature uses to control the strength of coupling of trillions of synapses, and how is this control related to learning? These are issues to be considered in Part I. Some answers are given. With each answer, further questions are raised. This is like peeling an onion. The deeper you go, the more of the unknown is exposed. There are no absolute answers. The % Hebbian-LMS algorithm is proposed as nature’s synaptic learning algorithm. It is a realization of the Hebbian learning rules bymeans of the LMS algorithm from the field of adaptive signal processing. Neurons whose synapses are controlled by % Hebbian-LMS can be connected in a network. Synaptic learning takes place independently with each neuron. Network inputs could come from sensors supplying visual, tactile, auditory, olfactory, etc. patterns. Connected in a network, Hebbian-LMS neurons perform clustering, automatically making connections and associations between objects observed in input patterns. Clustering seems to be a fundamental element of human thinking and learning.